Lydia Leong has elevated the discussion on the commoditization of cloud infrastructure in her response to my original post titled “Cloud Infrastructure: The Next Fat Dumb and Happy Pipe?” Her post, titled “Does cloud infrastructure become commoditized?“, rightly points out that added value has a life of its own that will tend to allow infrastructure providers to deliver value through customer service, good enough economies of scale, or other value-added capabilities that surround the infrastructure. She also rightly points out that it will take time for all this to mature and that means differentiation for quite some time. I think we are mostly in agreement although we differ perhaps in time-frame and aggressiveness of the commoditization. I argue that just as we don’t need 50 phone companies, we don’t need 50 cloud infrastructure giants. Certainly, Lydia is correct when she says there will be many players over the next 10 years. But the trend will be an explosion of infrastructure providers and then a decline over time as they begin to glut the market.
An old article published on the Internet a while back makes plain why this is so within an area like cloud infrastructure when it says “There is no formula for combating commoditization; how it is dealt with largely depends on the nature of the industry and the mode of competition therein. Commoditization is less likely to infect markets that require more capital investment to enter, such as heavy manufacturing. But even those industries are affected by burgeoning online business-to-business marketplaces. The capital investment required to enter into the modern information technology and computer software industries, meanwhile, is relatively small. As technology develops, it gets smaller all the time.” Read the complete article here.
And because of this (among other reasons I have mentioned in my post), I feel that it is only a matter of time for cloud infrastructure (in the public cloud) to commoditize as well. Then we will see rapid decline of the number of infrastructure providers. But whether you believe it will be 5-7 years, or 10-15, there will be a decline. And if you are in that market when the decline begins, the rug may be pulled out from under you faster that you ever imagined.
So, thanks, Lydia for elevating the discussion and clarifying the issues and time frame. But, let’s make a bet. I’ll place my wager on a cloud services provider that delivers value added on TOP of infrastructure (i.e. delivers business services or application services, or information services, or brokerage services, etc.) while you bet on a pure cloud infrastructure provider who delivers just an infrastructure platform. In ten years, lets count the percentage growth (positive or negative) in each area and we will see.
The winner buys because she/he will be flush with cash.
P.S. Here is an excerpt from an upcoming piece of research I am publishing on cloud brokerages, which is where a lot of money will be made in value-added services.
“The future of cloud computing will be permeated with the notion of brokers who broker the relationship between providers of cloud services and the service consumers. In this context, a broker might be software, appliances, platforms, or suites of technologies that provide enhancement of the base services that are available through the cloud. Enhancement will include managing access to the services, providing greater security, or even creating completely new services. The purpose of these brokers will be to add value to existing services, and to deliver new services and opportunities built and delivered on top of the old.”
The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.
Comments are closed
4 Comments
I’m not sure I 100% follow your thoughts on commoditization of cloud infrastructure. Are you saying that you think that clouds will use commodity hardware and software or are you saying that the clouds themselves will be like commodities?
If you’re saying its the first then that’s sort of equivalent to a the commodity (good-enough) components being the most effective because they’re cheaper. Couldn’t one easily disprove this by building a cheaper to use, more efficient cloud with proprietary components, but using the same interfaces as a cloud with commodity hardware and software? Clouds enable the users to distance themselves from hardware choices and focus on the APIs and machine images. If the interfaces were compatible would a customer care about the hardware or the price?
If you’re saying the second then do you see any differentiation between clouds. Couldn’t one develop an cloud optimized for high performance technical applications or with the added security feature needed for the healthcare?
These specialized clouds would conceivably allow for more market players as the requirements of the applications might actually require changes in the hardware and software. A cloud provider who went this route might be able to charge a different rate than a cloud provider using commodity components as a different level of service would be provided.
Anyway, I appreciate your thoughts on this.
Russ
Russ, the issue is that of commiditization of the business of being a cloud infrastrucutre provider. to sync up our language, I will say this. I believe there is one public cloud but many cloud providers who have different cloud platforms. so when we are talking about different cloud “platforms” (not different clouds what I mean is that delivering computer power or storage services, etc. will commoditize since those basic functions have little value to be added on top. There will be companies who add value like flicker does on top of a storagte service but within the provider, proprietary extensions to a cloud infrastructure service just makes it harder to interoperate with that servic. So, although it may take quite some time, the lower level cloud services like infrastruct will eventually commoditize as some degree of standardization happens. this happened with the Web in that being a Web Server provider became a bad business to be in if you wanted to generate revenue. some used CGI to differentiate site capabilities but this too was replaced by a more high-levelway of adding programmatic function to the web. As more players deliver cloud infrastructure, the common patterns of use will arise and standards will emerge. thus does NOT mean that all cloud services will be interoperable – only that in the long run, it won’t be a sustainable, differential blke revenue generating business for a lot of players. A smaller set of players will eventually win out.
Does this mean that cloud infrastructure providers will eventually end up on a cost-plus model, similar to existing (power/water) utilities?
Peter, that depends on how close they can stay to the ultimate consumers. the farther away from the ultimate consumer you are in a service Oriented model, the more likey it is that you will be at a level of cost recovery rather than value add. That is what happened to power, water, telecom, and a host of other basic infrastructure providers. those who add value (as Lydia suggests) will have to be closer to the ultimate consumer of the solutions that ride on top of them or the equation wkill just take longer to drive them into the commoditization. For example: In this chain of suppliers, who would YOU – the ultimate customer – care most about. IPTV and On Demand Movies to your home frovided by ATT-Uverse –> alcatel Lucent (pipe provider with value add). Alcatel is pretty critical and important, but most users don’t even know they are in the equation and the same will be true of enterprise users who pay for cloud services.